[00:00:00] Speaker 02:
this morning is 21-1917 Polaris Innovation versus Advanced Micro Devices.

[00:00:07] Speaker 02:
How do I say your name, counsel?

[00:00:09] Speaker 02:
Mr. Freitas, please proceed.

[00:00:13] Speaker 01:
May it please the court.

[00:00:15] Speaker 01:
The principal issue on Polaris' appeal involves the construction of the claim term checksum, which under the 526 patent is a type of check information.

[00:00:27] Speaker 01:
And under the proper construction of that term,

[00:00:29] Speaker 01:
The combination of the Liu and Iglesia references on which the board relied does not render the Claim 9 unpatentable.

[00:00:38] Speaker 01:
But there's another issue that does not depend strictly on the resolution of that base claim construction issue.

[00:00:46] Speaker 01:
That issue involves the Claim 9 limitation requiring that the processor be configured to determine the check sum based on the data bits and the first and second indicators.

[00:01:00] Speaker 01:
Much of the information on which Polaris relies and its position for claim construction and its position on the based on limitation is agreed or was determined by the board to be correct and isn't challenged by AMD.

[00:01:18] Speaker 01:
So the foundation for Polaris' argument is the teaching of the 526 patent on the differences among the various types of check information.

[00:01:27] Speaker 01:
That's where we begin.

[00:01:29] Speaker 01:
All appear to agree.

[00:01:30] Speaker 01:
that check information is the broadest category, and that there are subcategories of check information.

[00:01:37] Speaker 04:
So checksum's not the same as error correction codes.

[00:01:40] Speaker 01:
It's not, Your Honor.

[00:01:41] Speaker 01:
It's broader.

[00:01:43] Speaker 01:
I'm not sure whether it's broader.

[00:01:48] Speaker 01:
It's clearly different, because the claim language distinguishes between the two, that all of the subcategories are distinct.

[00:01:59] Speaker 01:
And by the way,

[00:02:00] Speaker 01:
Along the way in its analysis, the board pointed out that the applicant used his own lexicography in designing the claims in this package.

[00:02:15] Speaker 01:
So we don't have an overriding term of art, and we don't have a clear, comprehensive definition in extrinsic evidence.

[00:02:22] Speaker 01:
But there is some useful information in the extrinsic record, which all appear to agree is useful.

[00:02:31] Speaker 04:
In the Microsoft computer dictionary definition of checksum, the definition... I'm a little confused if you're arguing that the patentee acted as their own lexicographer while you're immediately jumping to extrinsic information.

[00:02:48] Speaker 04:
Usually when we're looking at coin terms of the patentee, we look at the patent and its specification because that's where you would find them defining their own terms.

[00:02:59] Speaker 01:
Your Honor, the reason why I'm turning to that is because I'm going to point to one specific aspect of the extrinsic record that is agreed.

[00:03:10] Speaker 01:
That's why.

[00:03:12] Speaker 01:
And by the way, Your Honor, the point that was made by the board about the lexicography, the point was that the Microsoft dictionary defines the term in a manner that excludes error correction.

[00:03:26] Speaker 01:
But under this pattern, it's quite clear

[00:03:28] Speaker 01:
that error correction is included.

[00:03:30] Speaker 04:
So that was the reason... What do you mean error correction is included in checksum?

[00:03:36] Speaker 01:
In check information.

[00:03:39] Speaker 01:
So a checksum can include error correction or error detection, but the point that was made... But it's not limited to error correction.

[00:03:50] Speaker 01:
It's not limited to error correction, nor is it limited to error detection, which the Microsoft definition suggests.

[00:03:59] Speaker 02:
But I didn't understand the district court's construction to limit it to anything.

[00:04:05] Speaker 02:
I understood the dispute to be here that you didn't want the construction to include simple parity bits, but the district court's construction included the ECC bits, parity bits, all different kinds of bits that could be part of the checksum.

[00:04:23] Speaker 01:
What Polaris specifically argued, Chief Judge Moore, is that

[00:04:28] Speaker 01:
The definition doesn't include bits that are calculated using non-overlapping inputs.

[00:04:37] Speaker 01:
That was the specific point.

[00:04:40] Speaker 04:
But the other argument, based on- Where in the specification does it say checksum doesn't include bits using non-overlapping?

[00:04:50] Speaker 01:
It doesn't say that in those words, judges.

[00:04:53] Speaker 00:
But your argument was less psychography, right?

[00:04:57] Speaker 01:
That's right, Judge Proost.

[00:04:58] Speaker 01:
What the patent shows, the patent uses the terms in a certain way.

[00:05:03] Speaker 01:
Equation 14 and other examples show only calculation using the overlapping inputs.

[00:05:12] Speaker 01:
And that's what we rely on, Judge Hughes.

[00:05:13] Speaker 01:
We don't have a distinct statement.

[00:05:17] Speaker 01:
What we have is a consistent use of the patent.

[00:05:20] Speaker 02:
I don't understand.

[00:05:21] Speaker 02:
You responded to Judge Hughes' comment or question about whether or not a checksum includes

[00:05:28] Speaker 02:
um, error correcting bits.

[00:05:30] Speaker 02:
And you said, what?

[00:05:32] Speaker 02:
You said it doesn't or it's not the same as, but it includes them, right?

[00:05:37] Speaker 02:
It includes error.

[00:05:38] Speaker 01:
So what, what we have to say on that, your honor is based on the claim language.

[00:05:45] Speaker 01:
So we have check information.

[00:05:47] Speaker 02:
Yes or no.

[00:05:48] Speaker 02:
Does a check sum include error correcting bits as one of what could be a type of bit within the check sum?

[00:05:55] Speaker 01:
Check information includes error-correcting bits.

[00:05:59] Speaker 01:
Check sum.

[00:06:00] Speaker 02:
What does check sum include?

[00:06:02] Speaker 01:
A check sum can include error-correcting bits, as claim five shows.

[00:06:08] Speaker 01:
Each check sum comprising error-correcting bits.

[00:06:11] Speaker 02:
Right.

[00:06:11] Speaker 02:
So if claim five says a check sum includes error-correcting bits, why doesn't that end your argument on the claim construction?

[00:06:21] Speaker 01:
It doesn't end our argument, because that doesn't resolve the question.

[00:06:25] Speaker 01:
of whether they're calculated using overlapping, non-overlapping bits.

[00:06:31] Speaker 01:
That issue is not resolved by that.

[00:06:40] Speaker 04:
So I'm really confused by that statement.

[00:06:44] Speaker 04:
Is error correction always using overlapping bits, or can it use non-overlapping bits?

[00:06:50] Speaker 01:
Error correction can use non-overlapping bits.

[00:06:54] Speaker 04:
So if checksum can include error correction bits, which I think you just said, and error correction bits can be non-overlapping, then why doesn't checksum include non-overlapping?

[00:07:10] Speaker 04:
Not bits, but bits produced by a non-overlapping method.

[00:07:14] Speaker 01:
Our argument, Your Honor, is based on the way

[00:07:20] Speaker 01:
Embodiments are consistently described in the patent.

[00:07:24] Speaker 04:
So you're asking us to rely on the embodiments rather than the claim language?

[00:07:29] Speaker 01:
On the consistent embodiments.

[00:07:30] Speaker 01:
That's right, Your Honor.

[00:07:31] Speaker 01:
Yes, we are.

[00:07:32] Speaker 04:
But we never do that.

[00:07:33] Speaker 04:
We've always said the claims aren't limited to their embodiments.

[00:07:36] Speaker 04:
If there's specific claim language, absent some other indication in the patent that it's confined to those embodiments.

[00:07:44] Speaker 04:
So your honor.

[00:07:46] Speaker 04:
Where's the language that said this patent only includes examples covered by these embodiments.

[00:07:53] Speaker 04:
I'm pretty sure it's not there because you never say that.

[00:07:55] Speaker 01:
It's not there.

[00:07:56] Speaker 01:
You're right.

[00:07:57] Speaker 04:
So let me turn to- Then how are we going to rely on the embodiments?

[00:08:02] Speaker 04:
Isn't that contrary to our basic principles of claim construction when you have general claim language that, as you just said, it could include error correction bids, that could include non-overlapping methods?

[00:08:15] Speaker 04:
You have to have a clear disclaimer or some other argument for why the specification excludes that.

[00:08:23] Speaker 01:
I understand what Your Honor is saying.

[00:08:27] Speaker 01:
Obviously, you're correct when you describe how things are generally done.

[00:08:32] Speaker 01:
So I'd like to turn to the other argument about the based on limitation.

[00:08:39] Speaker 01:
This is the one that we argued.

[00:08:41] Speaker 01:
The board said it wasn't contested, but it was.

[00:08:46] Speaker 01:
And AMD attempted to answer that in a way that isn't consistent with the patent.

[00:08:52] Speaker 01:
Now, we go back to the basics.

[00:08:54] Speaker 01:
Check information.

[00:08:56] Speaker 01:
it has various subcategories.

[00:08:59] Speaker 01:
One of them is a parity bit.

[00:09:01] Speaker 01:
But we know that a parity bit is not a checksum because the claims distinguish between checksums and parity bins.

[00:09:10] Speaker 02:
No, because a parity bit can be part of what creates a checksum.

[00:09:16] Speaker 02:
I mean, a parity bit is just a non-overlapping thing.

[00:09:20] Speaker 02:
You already said

[00:09:21] Speaker 02:
Checksum can include error-correcting bits, which include non-overlapping bits.

[00:09:26] Speaker 02:
And that's what a simple parity bit is.

[00:09:28] Speaker 02:
So, boom.

[00:09:29] Speaker 02:
I mean, you're right.

[00:09:30] Speaker 02:
They're not the same thing.

[00:09:31] Speaker 02:
A parity bit is a subset of what could be making up a checksum.

[00:09:36] Speaker 01:
Well, it could make up a checksum.

[00:09:38] Speaker 01:
But we know that definitionally, a parity bit is not a checksum because the claims distinguish between parity bits and checksums.

[00:09:47] Speaker 01:
I don't think that's disputed.

[00:09:48] Speaker 02:
They don't distinguish between them as though they're different items.

[00:09:50] Speaker 02:
They distinguish between them as though they're not, you know, completely overlapping Venn diagrams.

[00:09:55] Speaker 02:
But a parity bit can be a bit within a checksum.

[00:10:01] Speaker 01:
Can be a bit within a checksum but is not a checksum.

[00:10:05] Speaker 01:
So a single parity bit is not a checksum.

[00:10:08] Speaker 01:
The argument then becomes for AMD that when you have multiple parity bits, even though a single parity bit is not a checksum, the multiple bits become a checksum.

[00:10:20] Speaker 01:
And AMD argues that that is so because multiple parity bits can be used to calculate a value.

[00:10:28] Speaker 01:
This is where that aspect of the Microsoft dictionary comes in.

[00:10:33] Speaker 01:
A checksum is a calculated value.

[00:10:36] Speaker 01:
So AMD says when you have multiple parity bits, each individual bit not being a checksum, it is possible to calculate a value from the set.

[00:10:49] Speaker 01:
But what AMD cites to try to address the based on argument doesn't involve a value calculated from a set of parity bits.

[00:11:01] Speaker 01:
They're distinct bits.

[00:11:03] Speaker 01:
They are not used to calculate a value.

[00:11:06] Speaker 02:
What about claim E?

[00:11:07] Speaker 01:
Pardon me?

[00:11:08] Speaker 02:
What about claim E?

[00:11:10] Speaker 02:
Claim E. I think it directly refutes your proposed exclusion.

[00:11:23] Speaker 02:
Checksum C is calculated by taking component-wise modulo two sum of two inputs, a first checksum and a second checksum, and each component is a parity bit calculated from disjointed inputs.

[00:11:36] Speaker 02:
So why doesn't claim eight directly refute what you're saying right now?

[00:11:41] Speaker 01:
Well, I don't think it does, Your Honor, because those are checksums that are in claim eight.

[00:11:48] Speaker 01:
But it certainly doesn't address the argument I'm making on the based on limitation.

[00:11:54] Speaker 02:
But each component of C is a parity bit.

[00:11:57] Speaker 02:
That's all it is.

[00:12:02] Speaker 01:
In that case, Your Honor, where you have a value that's calculated from multiple parity bits, you have something that's different from what AMD is arguing to address the based on limitation.

[00:12:15] Speaker 01:
They're not showing multiple parity bits being used to calculate a value.

[00:12:20] Speaker 01:
They're showing multiple independent parity bits.

[00:12:25] Speaker 01:
If the court has no more questions, I'd like to reserve the balance of your time.

[00:12:28] Speaker 02:
Of course.

[00:12:31] Speaker 02:
Council, how do I pronounce your name?

[00:12:33] Speaker 02:
Ashish Kapadia.

[00:12:35] Speaker 02:
Mr. Kapadia, please proceed.

[00:12:40] Speaker 04:
Can you just clarify for me what you're appealing that you didn't win on?

[00:12:44] Speaker 04:
Let's just assume we don't think you have standing across appeal.

[00:12:48] Speaker 04:
claims you've won on.

[00:12:49] Speaker 04:
You can argue for alternative grounds of affirmance, but you don't have standing.

[00:12:53] Speaker 04:
What's left that you're appealing that you didn't win on?

[00:12:56] Speaker 03:
That would include claims 10, 11, 13, and 10, 11, and 13, which depend from, I believe, claim one.

[00:13:14] Speaker 04:
And that all depends on

[00:13:17] Speaker 04:
the combination of Iglesia and Sidhara.

[00:13:22] Speaker 04:
I'm not saying that.

[00:13:24] Speaker 04:
Well, last one right, but that's what we have to look to.

[00:13:27] Speaker 03:
That's correct.

[00:13:28] Speaker 03:
Those two references, your honor.

[00:13:29] Speaker 04:
And that doesn't even extend to the second ground of Iglesia, Sidhara, and Yuya?

[00:13:36] Speaker 04:
I don't see 10, 11, and 13.

[00:13:38] Speaker 03:
Yeah, that is correct.

[00:13:40] Speaker 03:
So for the second ground that adds Ouya, that would be claims 2, 3, 5, 14, 15, 17, 21, and 23.

[00:13:47] Speaker 04:
Wait, so you're cross-appealing those two, though, aren't you?

[00:13:49] Speaker 03:
That is correct.

[00:13:50] Speaker 04:
OK, so it's not just 10, 11, and 13.

[00:13:52] Speaker 03:
That's right.

[00:13:55] Speaker 04:
OK, go ahead.

[00:14:01] Speaker 03:
I'd like to start by talking about

[00:14:05] Speaker 03:
the appeal briefly with regards to the construction of the term checksum and the invalidity of claim nine.

[00:14:14] Speaker 03:
For the term checksum, the board properly rejected Polaris's attempt to read in a negative limitation into the claims.

[00:14:22] Speaker 03:
And they did so relying on four pieces of evidence, the intrinsic record, the dictionary definition that Polaris's expert cited to,

[00:14:29] Speaker 03:
AMD's expert, which provided context for that dictionary definition, and Polaris's own position in district court in which that dictionary definition supported their position, that checksum simply meant a calculated value that is used to test data for the presence of errors.

[00:14:46] Speaker 03:
There is no mention of checksum excluding multiple parity bits or multiple parity bits that are formed in a particular manner, for example, calculated from disjointed sets of inputs.

[00:14:58] Speaker 03:
So taken together, the board properly rejected it and had substantial evidence to support that.

[00:15:03] Speaker 03:
And Polaris hasn't shown that the board did not have substantial evidence to support its construction.

[00:15:09] Speaker 03:
It's also important to note that Polaris, before the board,

[00:15:13] Speaker 03:
indicated to the board that it is not a case of disavowal.

[00:15:16] Speaker 03:
So this is left either with the case of lexicography, which the intrinsic record does not support, or a situation in which the term is consistently used, which again is not supported.

[00:15:28] Speaker 03:
I'll point to Judge Moore's question with regards to claim eight.

[00:15:32] Speaker 03:
Claim eight of the patent and the associated disclosure does show that a checksum can be calculated using a modulo two operation,

[00:15:41] Speaker 03:
the very same operation that's used to calculate multiple parity bits.

[00:15:46] Speaker 03:
And that operation for claim 8 and the associated disclosure has two inputs that are not overlapping.

[00:15:53] Speaker 03:
So it specifically goes to the type of language that is in Polaris's negative limitation.

[00:16:01] Speaker 03:
In fact, if you were to adopt their construction, I believe claim 8 would be read out of the pack.

[00:16:07] Speaker 03:
On that issue, under the board's proper construction of the term,

[00:16:11] Speaker 03:
checksum, the board properly found that claim nine is invalid.

[00:16:16] Speaker 03:
If there are no further questions on the appeal, I'd like to turn to the cross appeal.

[00:16:23] Speaker 03:
Now, the board.

[00:16:25] Speaker 04:
Let me see if I, because I'm a little confused about what you're arguing.

[00:16:30] Speaker 04:
Let me make sure I understand it.

[00:16:31] Speaker 04:
Are you challenging both the substantial evidence on motivation, no motivation combined,

[00:16:41] Speaker 03:
wait what the two Iglesia with Sahara that way or are you not challenging that anymore and all you're saying is the board failed to look at it the other way starting with Sahara and then going to Iglesia it's the latter your honor the board did not address the second combination starting with Shridhara and a motivation to combine that with the Iglesia we are not appealing the first combination that is starting with the Iglesia the petition had

[00:17:09] Speaker 03:
two combinations in it.

[00:17:10] Speaker 00:
Well, you're right about that.

[00:17:11] Speaker 00:
But tell me what I'm missing, because the petition at 87 through 89 talks about these combinations.

[00:17:18] Speaker 00:
And as I'm reading it, you've got one paragraph on the first, second paragraph on the second, and then the three paragraphs, which are the meat of your argument, seem to be only with respect to the first and not the second.

[00:17:32] Speaker 03:
That is correct, Your Honor.

[00:17:33] Speaker 03:
The subsequent paragraphs deal with the first combination.

[00:17:37] Speaker 03:
The second combination that starts with Shidara.

[00:17:41] Speaker 00:
Your argument in your petition, at least, consists of one paragraph.

[00:17:45] Speaker 03:
That is correct, Your Honor.

[00:17:47] Speaker 00:
So how much was the board supposed to do in terms of addressing that issue?

[00:17:53] Speaker 03:
The board was required to address the combination as found in the original petition for that ground.

[00:18:00] Speaker 03:
Under SAS and subsequent federal circuit cases like FICOR, it is clear that the board must address all grounds.

[00:18:08] Speaker 03:
And that includes all combinations that relate to those grounds.

[00:18:12] Speaker 03:
in order to satisfy.

[00:18:13] Speaker 00:
But addressing the ground now that the board explicates the basis upon which it's reaching its decision.

[00:18:20] Speaker 00:
And if you effectively, in the petition, didn't make a fulsome argument or an argument that was comprehensible, what is the obligation of the board to address it consist of?

[00:18:30] Speaker 00:
To just say one sentence and we reject this too?

[00:18:35] Speaker 00:
Is that what was missing in the board's decision?

[00:18:38] Speaker 03:
No, Your Honor.

[00:18:38] Speaker 03:
What the board needed to do was address the combination meaningfully.

[00:18:42] Speaker 03:
For example, this court.

[00:18:44] Speaker 00:
Well, what is meaningfully?

[00:18:45] Speaker 00:
Is you addressed it in your petition, or more so?

[00:18:48] Speaker 03:
At least as meaningfully as is addressed in the petition.

[00:18:51] Speaker 03:
And the combination here, I'd note, is a very simple one, because it takes Iglesia's framework, its unified framework, and implements that using known hardware that's found using Shridhara's framework and implements it in Iglesia's hardware.

[00:19:07] Speaker 03:
Now both of these deal with the same concepts of error correction and data bus inversion.

[00:19:13] Speaker 02:
Where did you set forth any kind of distinct motivation to combine that began with Srihara rather than Iglesia?

[00:19:22] Speaker 03:
So Your Honor, that combination is found in the second paragraph.

[00:19:26] Speaker 04:
So am I right?

[00:19:28] Speaker 04:
We're looking at the arguments starting on page 377 of the appendix.

[00:19:33] Speaker 02:
377?

[00:19:36] Speaker 03:
87 87 your honor.

[00:19:39] Speaker 04:
Oh, I'm not I'm looking at the are they in different places?

[00:19:42] Speaker 04:
I don't know what this is.

[00:19:46] Speaker 04:
Well doesn't matter.

[00:19:47] Speaker 03:
I think they're the same So on

[00:20:03] Speaker 04:
I think you're in a brief.

[00:20:04] Speaker 04:
I'm in the brief.

[00:20:05] Speaker 04:
I think it's similar.

[00:20:06] Speaker 04:
But in any event, we're on page 87.

[00:20:11] Speaker 03:
On page 87 of the joint appendix.

[00:20:12] Speaker 03:
Sorry for interrupting.

[00:20:14] Speaker 02:
No, that's OK.

[00:20:15] Speaker 02:
Where is it that you set forth a distinct motivation if a person of skill in the art began with Srihara rather than Iglesia?

[00:20:26] Speaker 03:
So in the second sentence of the ultimate paragraph on page 87, it states, to the extent that Shridhara does not explicitly disclose a system with a processor and circuits, a person of skill and the art would have been motivated to look for ways to implement.

[00:20:42] Speaker 02:
That's saying a person of skill and the art would have looked for a way to do something.

[00:20:46] Speaker 02:
That's not a motivation to combine.

[00:20:48] Speaker 02:
That's not a specific argument.

[00:20:50] Speaker 02:
What is the motivation here?

[00:20:52] Speaker 02:
What is the motivation to make the combination that you've articulated?

[00:20:56] Speaker 02:
You know efficiency is it there's a there's a problem with the original reference that doesn't kind of that's that that concept is Conclusory that is not a specific motivation So the first reference should there are your honor.

[00:21:10] Speaker 02:
No, I don't want you to sit here and try to create a motivation now.

[00:21:13] Speaker ?:
I

[00:21:14] Speaker 02:
The problem I have is you didn't argue a separate motivation if you start with the one reference and go to the other.

[00:21:20] Speaker 02:
The only motivation you argued was with regard to the combination going the other direction, and the board fully addressed that.

[00:21:26] Speaker 02:
So if you didn't separately include a motivation that is distinct, if you were to start with Srihara and then go to Iglesia, then the board has already addressed in its entirety your

[00:21:41] Speaker 02:
ground because it rejected the only motivation you proffered with regard to combining these two references.

[00:21:48] Speaker 03:
If there is no motivation to combine, I agree the board doesn't need to address it.

[00:21:52] Speaker 03:
As in, if there is no combination effectively in the petition, then the board doesn't need to address it.

[00:21:56] Speaker 02:
Well, they already addressed the motivation to combine these two references.

[00:21:58] Speaker 02:
Did they not in the board's decision?

[00:22:01] Speaker 03:
They only addressed the motivation to combine starting with Eglacia.

[00:22:04] Speaker 02:
The motivation to combine Eglacia and Shrihara was already rejected.

[00:22:08] Speaker 02:
The motivation you proffered was rejected by the board as not sufficiently proven, correct?

[00:22:13] Speaker 03:
I would respectfully disagree that that's the only motivation.

[00:22:17] Speaker 04:
Well, first of all, you're trying to basically, after the fact, invent a new ground for

[00:22:24] Speaker 04:
in validity when the heading talks about Inglacian view of Siddhartha.

[00:22:30] Speaker 04:
Anyways, I mean, even if we allowed that, I mean, the board doesn't have to hunt and search for every single little possible combination you may be raising.

[00:22:40] Speaker 04:
They get to address what you raise.

[00:22:42] Speaker 04:
And what you raised is Iglesia in view of Sahara.

[00:22:46] Speaker 04:
And you explain why you would have used Sahara's stuff to fill out whatever.

[00:22:52] Speaker 04:
But do you explain you would start with Sahara and use other things in Iglesia, and that's how a person of skill would use that?

[00:23:02] Speaker 04:
Where does it say that here?

[00:23:03] Speaker 03:
So in this paragraph, Your Honor, it states, I'll go to the last sentence on page 87 of the Joint Appendix, that because

[00:23:12] Speaker 03:
Because Shridhara only discloses a framework.

[00:23:15] Speaker 03:
So it may be important here to distinguish what a framework is from actual hardware.

[00:23:20] Speaker 03:
A framework is simply just a concept.

[00:23:22] Speaker 03:
So it's a research paper.

[00:23:23] Speaker 03:
Shridhara is a research paper.

[00:23:25] Speaker 03:
It doesn't actually disclose a system, actual hardware.

[00:23:29] Speaker 03:
In order to implement that system to effectuate the teachings of Shridhara, you would need to look for

[00:23:35] Speaker 03:
systems that are in the art.

[00:23:37] Speaker 03:
That's what's described here.

[00:23:38] Speaker 03:
And the last sentence says, that would have led someone of skill to Iglesia.

[00:23:43] Speaker 03:
And why?

[00:23:43] Speaker 03:
It's because Iglesia describes a system architecture for performing the very functions that Shridhara's framework is designed to implement.

[00:23:54] Speaker 03:
And so that motivation is a simple one of providing that hardware.

[00:23:57] Speaker 03:
Now, in the petition, it repeatedly refers to Iglesia's hardware.

[00:24:02] Speaker 03:
For example, in the further preamble,

[00:24:05] Speaker 03:
to the extent it's limiting, Iglesia's hardware is pointed to for the overall system.

[00:24:12] Speaker 03:
And further, within claim one, there is a circuit that transmits information.

[00:24:17] Speaker 03:
That circuit in the petition specifically refers to the circuit that is disclosed in Iglesia.

[00:24:24] Speaker 03:
So what we have here is a motivation to combine and a framework for the first reference,

[00:24:31] Speaker 03:
and a way to implement that using the hardware for the second reference, ekklesia.

[00:24:36] Speaker 03:
And that is repeatedly referred to throughout the petition.

[00:24:42] Speaker 03:
Now, had the board considered this combination, the board would have found the motivation to combine.

[00:24:47] Speaker 03:
It's a simple combination, as I mentioned earlier.

[00:24:49] Speaker 02:
Just out of curiosity, what is a ground?

[00:24:54] Speaker 03:
A ground, Your Honor, includes the legal basis for the challenge.

[00:24:58] Speaker 03:
the references that are part of that, and the combinations thereof, as well as the claims that are challenged.

[00:25:04] Speaker 02:
I think I may be wrong, but I thought I remembered you responding to Judge Pritz's initial question by suggesting we raised two grounds in the petition, Iglesia in view of Srihara and Srihara

[00:25:18] Speaker 02:
in view of Iglesia, those are two separate grounds, correct?

[00:25:21] Speaker 02:
Because you're arguing two separate.

[00:25:23] Speaker 02:
According to you, there are separate motivation to combine.

[00:25:26] Speaker 02:
The motivation to combine is not the same for all of them.

[00:25:28] Speaker 02:
So would that be two separate grounds?

[00:25:31] Speaker 03:
That's one ground within the petition, Your Honor.

[00:25:33] Speaker 03:
They're both captioned under ground one, but they're two separate combinations using the same set of references.

[00:25:38] Speaker 02:
How can that be one ground?

[00:25:41] Speaker 02:
If you're telling me that there are different considerations that go into each of them, how could you put them under a single ground?

[00:25:49] Speaker 03:
A ground does not specifically delineate one direction, if you will, of a combination from another, but rather includes the references and the invalidity theories as an obviousness under section 103.

[00:26:03] Speaker 04:
So are you just, is what you're saying, your ground is you would have been motivated to combine these two references, and you could combine them both ways?

[00:26:15] Speaker 03:
That is correct, Judge.

[00:26:16] Speaker 00:
So you're not saying that the board didn't address one of your grounds.

[00:26:19] Speaker 00:
You're saying that we had several theories that we threw out there with respect to what would justify the ground raise.

[00:26:26] Speaker 00:
And the board was required to give a detailed explanation of those theories, even though they're essentially listed in your petition as saying whether you start from one or the other, you would have been motivated to come out.

[00:26:40] Speaker 03:
I would say that the extent theories includes combinations and the way in which the combination is cabined and motivation combined.

[00:26:50] Speaker 04:
Maybe they didn't address this second argument to your liking, but they start out their analysis in that section that they agree with the patent owner that you haven't shown that there is a motivation to combine these two teachings.

[00:27:04] Speaker 04:
They don't limit it to Iglesias over Siddhara.

[00:27:07] Speaker 04:
They say no motivation to combine these two teachings.

[00:27:10] Speaker 04:
Why isn't that enough?

[00:27:12] Speaker 03:
Because the board is required to reach all grounds.

[00:27:16] Speaker 04:
You just said these aren't separate grounds.

[00:27:18] Speaker 04:
The ground is Iglesias and Siddhara, and the board considered that combination.

[00:27:24] Speaker 04:
They gave fairly detailed explanations why they were rejecting it.

[00:27:30] Speaker 03:
The board is required under this court's decision in Donner Technologies to meaningfully address both the arguments and the evidence that's put forth in the petition.

[00:27:39] Speaker 03:
In this case, the board did not address that combination.

[00:27:42] Speaker 04:
That's just because we have to be sure that the board has looked at your arguments.

[00:27:46] Speaker 04:
It's clear they looked at your arguments, though.

[00:27:49] Speaker 04:
We don't grade their opinions for their thoroughness.

[00:27:53] Speaker 04:
It's just, is it sufficient to review under the APA for their conclusion?

[00:28:00] Speaker 03:
I see that my time is up.

[00:28:01] Speaker 03:
May I answer your question?

[00:28:04] Speaker 03:
Sure.

[00:28:05] Speaker 03:
Quickly.

[00:28:05] Speaker 03:
The board is required to address, to meaningfully address the arguments.

[00:28:09] Speaker 03:
If the board is simply just addressing the ground as a whole, when it comes to a combination that is insufficient in order to meet the requirements under the APA, and that's consistent with this court's cases.

[00:28:21] Speaker 02:
Thank you.

[00:28:23] Speaker 02:
All right.

[00:28:23] Speaker 02:
Mr. Freitas, you have some available time.

[00:28:28] Speaker 01:
I'd like to address my rebuttal argument to the distinct argument that's made beginning on page 29 of our brief regarding the based on limitation.

[00:28:43] Speaker 01:
My friend mentioned claim eight and said that it answers various things, but it doesn't answer this issue.

[00:28:50] Speaker 01:
Under claim 11, the check information must be a parity bit.

[00:28:55] Speaker 01:
under other claims, the check information must be a checksum.

[00:29:00] Speaker 01:
So those are different things.

[00:29:01] Speaker 01:
Everyone agrees that a single parity bid is not a checksum.

[00:29:06] Speaker 00:
Well, they can be distinct concepts, but that doesn't answer the question of whether or not a checksum can include a parity bid, right?

[00:29:14] Speaker 01:
It doesn't answer that question, Judge Prost.

[00:29:17] Speaker 01:
However, a single parity bid is not a checksum.

[00:29:21] Speaker 01:
A checksum can include a parity bid.

[00:29:25] Speaker 01:
The problem that exists, and the problem that we argue beginning on page 29 of our brief, is that given that a single parity bit isn't a checksum, to make multiple parity bits a checksum, one has to demonstrate a calculated value associated with the multiple bits.

[00:29:47] Speaker 01:
If they're simply standing by themselves and no value is calculated from them, they don't amount to a checksum.

[00:29:55] Speaker 01:
And that's the point that AMD hasn't addressed.

[00:29:59] Speaker 01:
AMD argues, and the board appears to have decided, that simply transitioning from a single bit to multiple bits transforms the non-check sum single bit into a check sum.

[00:30:16] Speaker 01:
Theoretically, that could happen.

[00:30:18] Speaker 01:
And the arguments that are being made explain how multiple parity bits could comprise a check sum.

[00:30:25] Speaker 01:
But in the example on which AMD relies, the individual parity bits are individual parity bits.

[00:30:34] Speaker 01:
They are not being used to calculate a value.

[00:30:37] Speaker 01:
And that's the point that they haven't answered and that the board didn't address.

[00:30:42] Speaker 01:
The board did not address Polaris's argument regarding the based on limitation, as we explained beginning on page 29 of our brief.

[00:30:54] Speaker 01:
The issue hasn't been decided.

[00:30:56] Speaker 01:
AMD's attempt to show that what Polaris has to say is incorrect failed.

[00:31:02] Speaker 01:
And that's where we are on that issue.

[00:31:05] Speaker 01:
And by the way, this is a slightly different point, but Dr. Zierbilski pointed out that separate checksums remain separate checksums.

[00:31:20] Speaker 01:
We submit there's no basis.

[00:31:23] Speaker 01:
for distinguishing parity bits on that score.

[00:31:27] Speaker 02:
OK, thank you.

[00:31:29] Speaker 02:
Mr. Kredos did not address the cross-appeal, so that concludes all of the argument in this case.

[00:31:35] Speaker 02:
This case is taken to submission.